And before you know it, the 9th PHUSE/FDA Computational Science Symposium, better known as the CSS, is over. Holding this working meeting virtually over three days was a first for PHUSE – and it was a huge success. For those who attended, presentations are now available on the ON24 Engagement Hub.
Over the course of the event, we heard updates from the FDA, watched 17 poster presentations, held 2 panel discussions and an interactive Hackathon, not to mention over 60 Working Groups breakout sessions and a Synthetic Data Workshop. There were also on-demand recordings from the Safety Analytics Working Group and PHUSE Education.
The PHUSE Working Groups and projects are made up of many knowledgeable and enthusiastic volunteers who make significant contributions throughout the year. We are very grateful for the efforts of all of those involved to advance the PHUSE Working Groups’ mission to address unmet computational science needs in support of health product development and regulatory review, ultimately bringing safe and effective medical products to those who need them.
We would like to thank everyone involved for their time and dedication in contributing to what was an extremely successful event. A full summary of the CSS will be presented by Chris Price, PHUSE Working Groups Director, at our October Webinar Wednesday, on the 28th at 15:00 (GMT).
Prize Winners
CSS 2020 Plane Builder Award Winners
Each year at the CSS, PHUSE recognises an individual in each Working Group who has been a key contributor over the previous 12 months. PHUSE would like to acknowledge this year’s CSS 2020 Plane Builder Award Winners and would like to thank them for their continued contribution throughout the past year.
Nancy Brucken – Standard Analyses & Code Sharing
Nancy has been very active in multiple projects. She has broad experience, having worked with multiple sponsors and therapeutic areas. Her knowledge and interest span across the entire data flow from collection through to analysis. Having such expertise has been invaluable to the project teams; she’s a true advocate for us.
Bob Friedman – Nonclinical Topics
Bob is a highly engaged PHUSE participant, with numerous valuable contributions to the Nonclinical Topics Working Group throughout the years and careful integrity to the PHUSE mission.
His knowledge sharing and computational expertise has enabled strong collaborations across not only Nonclinical Topics but also across the other PHUSE Working Groups. Bob’s contributions include delivery of the SEND Data Factory, Automated Creation of TS Domain to satisfy a Technical Rejection Criteria, coordination of Nonclinical Topics participation at the PHUSE US Connect and the annual SEND Implementation Survey, as well as many posters representing the work of his team members for the CSS. Bob’s valuable expertise in the field of nonclinical computational science has made and continues to make a difference for PHUSE!
Kiran Kundarapu – Optimizing the Use of Data Standards
Kiran has been actively involved in multiple Optimizing the Use of Data Standards projects, including the publication of a final deliverable on Event Adjudication, a project in which he led.
Kiran contributes to the SDTM ADaM Implementation FAQ forum and the Industry Experiences Submitting Standardised Study Data to Regulatory Authorities project and has presented at PHUSE and FDA events. Kiran and his team are working towards the publication of their next white paper on the FDA-PMDA Study Data Submission Distinctions. He is Project Co-Lead for Clinical Integrated Study Data and Analysis Data Reviewer’s Guide. Kiran has gone above and beyond in his contributions to PHUSE.
Anne Cutting – Data Transparency
Anne joined the Data Transparency Working Group about two and a half years ago. She is a Data Sharing Expert at GSK and was previously an important contributor to the TransCelerate Clinical Data Transparency workstream, co-leading the creation of its White Paper on the protection of personal data in clinical documents.
Anne led the mammoth effort for PHUSE to create a Best Practices Guide to complement the previously published white paper considering the multi-faceted area of clinical trial disclosure and data sharing. The result was a comprehensive manual addressing all pillars of data disclosure and sharing. The project to create this guide spanned over one and a half years and the 60-page document that was created has been unparalleled in the industry in terms of scope and practical content. Anne led a very active sub-team, who worked extremely hard through numerous rounds of reviews to ensure consistency and that all aspects were appropriately covered. Anne was also an important contributor to the initial White Paper that was the starting point for this Best Practices Guide.
David Hood – Emerging Trends & Technologies
Since its inception, David has been co-leading the PHUSE Real World Evidence project, focusing on establishing guidelines and garnering experiences in how real-world evidence can be incorporated into clinical trials.
Together with the RWE team, David organised the development of their first white paper, Basic Considerations for the Use of RWE, and continues to take a lead as the project moves to its next phase in this important new area. Congratulations, David!
Poster Winners
The live Poster Session took place on the second day of the virtual conference, and as with every year, we recognised the contributions made. Thank you to all that contributed, and many congratulations to our Poster Winners.
FDA – Jack Slattery & Matt Wiley, FDA
This poster provides an overview of the FDA’s comprehensive approach to end-to-end data management. OCS Data Central integrates a database repository, data management microservices, efficient workflow automation, and an integration with data quality that allows efficient data connection to the FDA’s tools and services. The authors present a compelling view of the FDA’s data management infrastructure that allows for efficient access to quality study data. The poster illustrates the FDA’s commitment to applying novel approaches to improve the regulatory drug review process.
Industry – Kunaal Madhavan & Laura Kaufman, PointCross
The poster provided great insight into the reconciliation process of SEND datasets with the existing Study Report. The authors demonstrated the use of automated “reconcilers” to read existing study reports in machine-readable columnar format and compare it with the retrospectively generated SEND datasets. They were able to lay out the process to demonstrate to the reviewers that the SEND datasets accurately represent data in the study reports. This poster will provide a great reference to the industry looking to make advancement in this topic.
Collaboration – Xiangyun Wang, Genentech & Zachary Skrivanek, Eli Lilly
PP16 Best Practice for Interactive Analysis for Submission - A PHUSE White Paper
This poster provided insight into an impactful white paper from a PHUSE project team. The team explored the benefits and technical feasibility of interactive analysis for regulatory deliverables. Noting the FDA’s awareness of slowing drug development when relaying a digitalised paper, the project team developed a Proof of Concept and a Pilot study to assess several approaches on how to share interactive tools with health authorities. Feedback from the FDA found that interactive analysis was useful. This also led to new ways to modernise existing eCTD processes. Future endeavours include connecting to other initiatives and to creating a common user interface. This poster is an excellent example of how industry collaboration can provide efficiencies in our drug review process.
Hackathon Winners
Due to the movement within the industry, as well as the exciting restructuring of PHUSE’s Working Groups – and therefore the inclusion of more technical topics – 2020 was the year we re-introduced the Hackathon. This was event-long, over the course of the CSS. Many congratulations to our winners for the Streams listed below.
Qualification Stream Award – Frank Menius, YPrime, Srinivas Veeragoni, Bayer
The Qualification Stream discussed important questions around the use of open source within regulatory environments. How can we create agile processes for adopt and qualify packages, while staying user-friendly and unambiguous? We capitalised on work done in the R Validation Hub white paper and evaluated a risk-based approach, ultimately creating an intuitive process flow that helps companies establish a decision tree around a process for incorporating open-source packages in the qualified environments.
ADaM Stream Award – Alexandra Papadopoulou, & Katherine Wilde, Janssen Research & Development
The ADaM Stream explored producing ADaM datasets, using synthetic data and specifications provided by Janssen Research & Development. We had interesting results on the R side, working through the challenges of CDISC dataset development. This revealed some challenges that may need to be overcome when using a language like R, such as the verbosity of messaging in R and tricky areas like working with dates. On the Python side, a different approach was taken. We explored how automation could be introduced to areas like ADaM, and how code can be made much more reusable.
TLF Stream Award – Bhavin Solanki, Bayer
The TFL Stream was fruitful as well. R code was written to produce an adverse events summary table. We also explored some specific challenges, capitalising on the PHUSE Analysis and Displays Associated with Adverse Events White Paper. We produced an instructional document on how to properly calculate study-size adjusted percentages using R. Additionally, we went through the exercise of vetting R packages for statistical results, as we explored how we wanted to use the Mantel-Haenszel method to calculate risk difference. Furthermore, we produced a Jupyter Notebook that outlines the data preparation steps necessary to produce an adverse events table in Python, which is an area that has been much less explored.
Hackathon Summary
This year we explored the ever-expanding clinical data scientist toolkit and what both established programming languages and the world of open-source languages like R and Python have to offer us and how we can apply these tools to our daily deliverables. This year’s Hackathon was broken into three Streams, where we explored producing ADaM datasets, generating data displays, and explored the tough questions around environment qualification.
We’d like to extend a huge thank you to RStudio PBC for allowing us to use their RTT environment. This gave us access to the RStudio environment for R development, and Jupyter Lab/Jupyter Notebooks for Python development. This allowed us to focus on the challenges we wanted to address in the Hackathon and less on technology. Additionally, we’d like to thank José Lacal for getting us equipped with the new PHUSE Open Data Repository (PODR), centralising all the data for the Hackathon in one place. In fact, one of the outcomes of the Hackathon was an R Shiny application to explore data in the PODR.
Thank you to all of the participants in the Hackathon this year! Our discoveries not only provide new informative resources to the PHUSE community but also open the doors for new challenges that PHUSE can take on in 2021. You can find all of the content produced in the Hackathon this year on GitHub.